Muscular Injury Is a Risk Factor for Post-Entire Circumferential Esophageal Endoscopic Submucosal Dissection Stricture

Endoscopic submucosal dissection (ESD) is a widely used treatment for gastrointestinal neoplasms, including esophageal cancer. With recent improvements in endoscopic techniques and strategies, ESD has become a feasible and effective approach for treating entire circumferential s involving more than three-fourths of the circumference of the esophagus, but are risk factors for post-ESD strictures. In particular, defects in the entire circumferential esophageal mucosa frequently result in post-ESD stricture [1, 2]. Although local injection or oral administration of steroids is commonly used to prevent post-ESD strictures [3], the relevance of strictures after entire circumferential ESD (EC-E-ESD) remains high [4].

Endoscopic balloon dilation (EBD) is a common salvage treatment for post-ESD esophageal strictures; however, frequent EBD is sometimes needed [5]. Refractory esophageal stricture requiring repeated EBD is a complication that significantly diminishes a patient’s quality of life (QOL) and increases healthcare costs. Therefore, elucidating the risk factors for refractory strictures following entire circumferential esophageal ESD (EC-E-ESD) is essential. Lesions above 5 cm have been reported as a potential risk factor for refractory strictures in EC-E-ESD. However, few studies have explored the risk factors for EC-E-ESD refractory strictures, and the existing data remain limited by small sample sizes [6, 7]. Therefore, this study aimed to clarify the risk factors for refractory strictures after EC-E-ESD.

Between July 2013 and September 2023, 53 patients underwent EC-E-ESD at the Yokohama City University Medical Center. The inclusion criteria for ESD in these patients were as follows: (1) depth of invasion of the lesion limited to the submucosal layer (<200 μm) as determined by pretreatment endoscopic findings, (2) absence of lymph node or distant metastases on computed tomography, and (3) written informed consent obtained before treatment. Patients who could not be followed up for at least 6 months and those who received additional chemoradiotherapy (CRT) after ESD were excluded from this study. Therefore, 49 patients were included and classified into two groups: refractory and non-refractory postoperative stricture groups. The study was conducted in compliance with the Declaration of Helsinki and received approval from the Yokohama City University Research Ethics Committee (IRB Registration No: F230700035). Since this study was a retrospective observational study, informed consent was obtained through an opt-out form on our institution’s website.

ESD was performed by experienced endoscopists under sedation with propofol or general anesthesia. The endoscopists involved in the cases examined in this study had a median experience of 50 esophageal ESD procedures. A single-channel upper gastrointestinal endoscope (GIF Q260J; Olympus, Tokyo, Japan) equipped with an electrosurgical unit (VIO300D; ERBE, Tubingen, Germany) was used. Tumor outlines were identified using iodine staining, and marker dots were placed on the anal and oral sides of the tumor margins using a dual knife (KD-650L; Olympus, Tokyo, Japan). Local injections were administered near the markings, and a mucosal incision was made using a dual knife in the VIO300D ENDOCUT mode. The injection solution comprised 20 mL of 10% glycerin, 20 mL of 0.4% sodium hyaluronate solution (MucoUp; Johnson and Johnson, Tokyo, Japan), and 0.2 mg of adrenaline. Endoscopic submucosal tunnel dissection (ESTD) and clip line traction were utilized to address entire circumferential lesions [8, 9]. Two submucosal tunnels were created on the proximal side using the ENDOCUT or coagulation mode. These tunnels were connected via dissection using an IT knife nano (KD-612; Olympus, Tokyo, Japan). During dissection, a clip with a thread was attached to the mouth of the tumor to ensure counter traction [10, 11]. Following dissection, the exposed vessels were cauterized using hemostatic forceps (Coagrasper; Olympus, Tokyo, Japan). After the procedure, plain chest radiography or chest computed tomography was performed to check for perforations.

At our center, steroid injection followed by oral steroid therapy was administered to all patients with post-ESD mucosal defects involving the entire circumference of the esophagus. The steroid injections comprised undiluted triamcinolone acetonide (Kenacort, 50 mg/5 mL; Bristol Myers Squibb, Tokyo, Japan). The solution was injected evenly into the residual submucosa of the ulcer base using a 25-gauge, 1.8-mm needle (TOP Corporation, Tokyo, Japan). The decision to administer 50 or 100 mg of triamcinolone acetonide was left to the operator’s discretion. As an oral steroid, prednisolone was initiated at 30 mg/day on the second day after ESD. The dosage was tapered gradually (30, 25, 20, 15, 10, and 5 mg) over 14-day intervals and discontinued 12 weeks later.

An initial endoscopic examination was planned approximately 14 days after ESD to evaluate the patient’s post-ESD ulcer and stricture status. Subsequent endoscopic examinations were performed regularly until the ulcer healed. The intervals of endoscopic examinations were adjusted every 7–28 days, depending on the patient’s condition. In cases of esophageal strictures, EBD was performed using an esophageal balloon dilation catheter (CRE Fixed Wire 10–12 mm/12–15 mm/15–18 mm; Boston Scientific Co., Boston, MA, USA). EBD was carried out using direct visualization. Triamcinolone acetonide was injected after EBD as appropriate, at the operator’s discretion. The EBD procedure was repeated until stricture improvement was achieved.

Post-ESD stricture was defined as the inability to pass a 9.9-mm diameter endoscope (GIF Q260J; Olympus, Tokyo, Japan) through the esophagus. Release of esophageal stricture was defined as the ability to pass a gastroscope without EBD during endoscopy, followed by an EBD-free period of at least 4 weeks. Refractory stricture was defined as the requirement for six or more EBD procedures before achieving the release of esophageal stricture. Non-refractory stricture was defined as requiring fewer than six EBD procedures before the release of esophageal stricture or absence of stricture altogether [12]. Muscular injury was defined as grades 3–4 according to Xu’s criteria [13] (shown in Table 1; Fig. 1). Endoscopic findings of muscular injuries were retrospectively analyzed using still endoscopic images by two experienced endoscopists (D.A. and R.K.) who were blinded to patient characteristics, tumor features, and histological diagnosis. For discordant cases, consensus was obtained after discussion. Mucosal defect length was defined as the length of the resected specimen, measured from the oral end to the anal end.

Statistical analyses were performed using JMP Pro 17.0 (SAS Institute, Cary, NC, USA). Continuous variables were reported as medians and ranges, whereas categorical data were presented as frequency counts and percentages. Between-group comparisons were performed using Fisher’s exact test, the χ² test, or Wilcoxon’s rank-sum test, as appropriate. Logistic regression analysis was used to identify independent predictors. For multivariable analysis, these factors were selected: muscular injury (grade 0/1/2 or 3/4), mucosal defect length (<50 or ≥50 mm), lesion location, and invasive depth. Odds ratios and 95% confidence intervals were calculated using the multivariable models. Statistical significance was set as a p value <0.05.

EC-E-ESD was performed on 53 patients. Two patients were excluded due to death for reasons unrelated to ESD and inability to follow up for more than 6 months. Additionally, 2 patients were excluded because they underwent non-curative resections and received additional CRT. Ultimately, 49 patients were included in this analysis, with 25 lesions in the refractory stricture group and 24 lesions in the non-refractory stricture group (shown in Fig. 2). Regarding patient demographics, the study included 37 men, with a median age of 71 years (range, 49–88). Regarding tumor characteristics, the middle thoracic region was the most common lesion location (47%), and pathological diagnosis revealed squamous cell carcinoma in 44 lesions (90%). Of the 49 lesions, 34 (69%) were histologically shown, in terms of depth invasion, to be confined within the lamina propria mucosal layer, while 13 (26%) involved a deeper layer than the muscularis mucosa. Regarding the treatment factors, the median procedure time was 107 min (range, 28–270 min), the median mucosal defect length was 50 mm (20–100 mm), and 15 patients (31%) had grades 3–4 muscular injury (shown in Table 2). In all cases, the esophageal stricture healed with EBD alone.

Refractory strictures were diagnosed in 25 patients (51%). A comparison between the refractory and non-refractory groups showed that the refractory group had a significantly higher percentage of muscular injury (52% vs. 8%, p = 0.002) and mucosal defect length ≥50 mm (68% vs. 37%, p = 0.047) (shown in Table 3). When comparing the rate of muscular injury grades 0 and 1–2 between the refractory and non-refractory stricture groups, no significant differences were observed (shown in Table 4). We performed a multivariate analysis of factors with significant differences in comparison between the two groups, lesion location, and invasive depth. The results indicated that muscular injury was an independent factor for post-EC-E-ESD stricture (odds ratio of 16.2 [95% confidence interval: 2.04–129.1]) (shown in Table 5). No significant differences were observed between the two groups regarding other factors.

In the refractory stricture group, all patients had their stricture corrected by EBD alone, and no other additional procedures were attempted. The median number of EBD sessions was 12 (range, 6–54). The median period until stricture was removed was 147 days (69–935 days).

This study demonstrated that a significantly higher proportion of patients with a muscular injury and mucosal defect length of ≥50 mm developed refractory esophageal stricture. Furthermore, multivariate logistic regression analysis of these factors, as well as lesion location and invasive depth – which have been previously reported as risk factors for post-ESD stricture – showed that muscle injury was an independent risk factor for refractory stricture after EC-E-ESD [2, 14, 15]. Therefore, avoiding deep muscular injury during EC-E-ESD is considered a crucial method for preventing refractory strictures.

In EC-E-ESD, even if the tumor is entirely circumferential, surgery or CRT can be avoided if the depth of invasion remains within the muscularis mucosa. This can contribute to maintaining the patient’s QOL. On the other hand, the likelihood of developing refractory strictures is high, in which case the patient’s QOL is significantly reduced. Additionally, previous studies have demonstrated that muscular injury and mucosal defects above 5 cm result in refractory strictures requiring multiple EBDs [6, 7]. Honda et al. [16] reported the healing process of post-endoscopic mucosal resection ulcers in a canine model. The authors suggested that atrophy of the muscularis propria during the healing process is associated with strictures following esophageal endoscopic resection. Nonaka et al. [17] studied the healing process of esophageal post-ESD ulcers using localized steroids in a porcine model, finding that the proliferation of spindle-shaped myofibroblasts during ulcer healing contributed to esophageal stricture. These findings suggest that fibrosis of the muscle layer is associated with esophageal stricture and that muscular injury may contribute to esophageal stricture. Xu et al. [13] concluded that deep muscle injury is a risk factor for strictures. Geng et al. [18] concluded that muscle layer injury was an independent risk factor after esophageal ESD, with deep or extensive muscle layer injury being associated with postoperative stricture. Our study showed a significantly higher rate of refractory stricture in grades 3–4, i.e., deep muscular injury cases, suggesting a similar conclusion for entire circumferential lesions. Furthermore, no significant difference was observed between muscular injury grades 0 and 1–2, suggesting that shallow muscular injury is not a risk factor for refractory stricture.

Additionally, localized steroid injections for post-ESD esophageal ulcers have been reported as an effective method for stricture prevention [19]. Since steroids are injected into the submucosal layer, they cannot reach areas where muscular injury has occurred. Therefore, we believe that the ineffectiveness of local steroid injection, coupled with the fibrosis of the muscle layer mentioned earlier, explains why muscular injury is a risk factor for refractory strictures. In this study, local steroid injections and oral steroids were used together as a prophylactic treatment for strictures. However, this combination therapy is insufficient to prevent strictures in cases of muscular injury.

Therefore, additional prophylactic measures should be considered for cases involving muscular injury. Various methods, such as polyglycolic acid sheets [20], combined localized steroid injection after EBD [21], transplantation of esophageal mucosa [22], prolonged duration of oral steroid administration [23], and oral steroid gel [24], have been reported as additional treatments following muscular injury. While most of these treatments have not proven useful for entire circumferential lesions, they may be useful in preventing strictures in cases of muscular injury. Further research and development in this area is expected.

Esophageal ESD is considered a challenging procedure due to the narrow lumen, thin walls, and difficulty in achieving adequate traction. The inevitably larger resection area required for an entire circumferential lesion complicates this procedure and increases the risk of muscular injury. To minimize muscular injury, specialized treatment methods and devices can be effective. In this study, ESD was performed using ESTD and clip line traction. ESTD, a technique for large lesions in which a submucosal tunnel is created, has been reported as a safe and fast technique for ESD [8, 9]. A meta-analysis of ESTD indicated a trend toward reducing the incidence of muscular injury, although consensus on this issue has not been fully reached [25]. Clip line traction has also been reported to be useful for esophageal ESD [26]. Kaku et al. [27] developed a new traction device, the Endo Trac. This device can adjust the direction of traction, making visualization of the submucosa easier. In the context of EC-E-ESD, this device is likely to be effective and may be a useful means of avoiding muscular injury. A meta-analysis comparing ESD with scissor-type knives and non-scissor-type knives reported that scissor-type knives were associated with a reduced incidence of perforation [28]. Scissor-type knives are particularly useful for precise dissection, minimizing the risk of damaging the muscle layer and thus potentially reducing the incidence of muscle injury. Additionally, the depth of dissection plays a key role in avoiding muscular damage. It is essential to focus on dissecting within the middle submucosal layer, ensuring sufficient submucosal tissue remains intact to protect the underlying muscle. Traction should also be applied in a direction that maintains an optimal visualization of the submucosal plane, avoiding over-tension that could distort the layers and increase the risk of muscular injury. Adjusting clip placement and ensuring that traction is gentle yet sufficient for proper submucosal exposure can prevent excessive stress on the dissection area. This technique can significantly reduce the risk of unintended muscular injury and improve procedural outcomes. Furthermore, the choice of high-frequency mode may also play a significant role in preventing muscular injury. Previous reports suggest that the ENDOCUT mode may help prevent postoperative fibrosis and stricture [29, 30]. In this study, the distinction between ENDOCUT mode and coagulation mode was not systematically recorded, and therefore, a detailed analysis of their respective impacts was not possible. Nonetheless, dissection using ENDOCUT mode may prove useful in minimizing stricture development. None of the techniques mentioned have been established as definitive methods for preventing muscular injury. Further improvements in techniques to safely perform EC-E-ESD are anticipated.

This study has some limitations. First, this was a retrospective, small-scale study conducted at a single center. Thus, further case accumulation and multicenter studies are required in the future. Second, muscular injury grading was performed by endoscopists, with multiple specialists being involved in the diagnosis; however, some bias may still have occurred. Third, the definition of stricture in this study did not include symptoms. It is possible that asymptomatic, clinically unproblematic cases were categorized as stricture cases solely because they could not accommodate a gastroscope.

Muscular injury is an independent risk factor for refractory strictures following EC-E-ESD. Avoiding muscle layer damage during ESD and adequate steroid injections may help prevent strictures. Currently, there are no established methods for preventing strictures other than intravenous steroid injections and oral steroids. Additionally, specific risk factors associated with refractory strictures have not yet been elucidated. Appropriately clarifying these gaps could allow for EC-E-ESD to be performed without causing refractory strictures, potentially expanding the indications for ESD.

This study protocol was reviewed and approved by the Yokohama City University Research Ethics Committee (Approval No. F230700035). All patients provided written informed consent for the procedure. Since this was a retrospective observational study, informed consent was obtained through an opt-out form on our institution’s website. Opt-out informed consent protocol was used for use of participant data for research purposes. This consent procedure was reviewed and approved by Yokohama City University Research Ethics Committee, Approval No. F230700035, date of decision May 13, 2024.

The authors declare no conflicts of interest.

This study was not supported by any sponsors or funders.

D.A., K.H., and R.K. led and designed the study. D.A., K.H., R.A., Y.O., A.S., M.N., R.K., and C.S. performed the endoscopic examination. D.A. wrote the manuscript and contributed to the statistical analysis. D.A. and R.K. contributed to data collection. K.H., R.A., Y.O., A.S., M.N., R.K., C.S., and S.M. critically reviewed the manuscript and contributed to the discussions. All authors read and approved the final manuscript.

The data supporting the findings of this study are not publicly available due to containing information that could compromise the privacy of research participants. However, they are available from the corresponding author upon reasonable request.